Reasoning, fast and slow: How noncognitive factors may alter the ability-speed relationship

Abstract

Although not unequivocal, a general viewpoint based on earlier research is that the ability-speed relationship in reasoning tasks is likely to be zero or slightly positive (Carroll, 1993; Kyllonen, 1985). Yet, more recent studies (Goldhammer & Klein Entink, 2011; Scherer, Greiff, & Hautamäki, 2015; Shaw, Oswald, Elizondo, & Wadlington, 2014) adopting the conjoint item response theory (CIRT) modeling approach (van der Linden, 2006, 2007) have found this relationship to be negative and moderate-to-large in size. Attempting to address such mixed findings, the current article proposes and examines the moderating effects of test situation and personality on the exhibited ability-speed relationship possibly via influencing test takers' choices of speed-accuracy tradeoff. Based on a sample of N = 300 working adults who completed a reasoning test and a personality assessment in a high-stakes selection context, we modeled item responses and response times as well as two personality traits (Conscientiousness and Neuroticism) in CIRT. In line with the overall conclusion by Carroll (1993), our results revealed a nearly zero ability-speed correlation. Comparing this finding to the negative correlations found in other CIRT studies, we contend that these negative relationships were likely due to low test-taking motivation in low-stakes contexts and that test situations matter in intelligence testing. Additionally, Conscientiousness and Neuroticism were found to be negatively related to speed but not ability on the test. Study implications, limitations, and future research needs are discussed.